Phonetics Lab researchers Annabelle di Lustro, Fengyue (Lisa) Zhao, and Chloe Kwon presented posters and papers at the 190th Meeting the Acoustical Society of America, held May 11-15, 2026 in Philadelphia, PA.

Their presentation and poster titles were:  

Laryngeal devoicing gestures observed using transverse-oriented ultrasound, by Fengyue (Lisa) Zhao and D4. Sam Tilsen (Cornell University) and Dr. Mark Tiede (Yale University)
Presented at the Speech Communication: Production & Articulation Poster Session (which was chaired by Phonetics Lab Alumnus Dr. Margaret Renwick of Johns Hopkins University)

Laryngeal elevation in native and non-native lexical tone production, by Fengyue (Lisa) Zhao and Dr. Sam Tilsen (Cornell University) and Dr. Mark Tiede (Yale University)
Presented at the Speech Communication: Acoustics Voice Quality and Prosody Poster Session.

Prosodically-conditioned temporal variation in whispered versus normal speech, by Annabelle Rose di Lustro and Dr. Sam Tilsen (Cornell University)
Presented at the Speech Communication: Acoustics Voice Quality and Prosody Poster Session.

Temporal localization of morphologically-conditioned stop categories in Korean: A computational approach, by Chloe Kwon and Fengyue (Lisa) Zhao (Cornell University) - Presented as a Contributed Paper at the Computational Acoustics: General Topics in Computational Acoustics Session.

On Friday, May 8, 2026, PhD Candidate John Starr successfully defended his dissertation:  Dynamic Interaction of Linguistic Streams During Processing - see the abstract below.

John is shown here with his dissertation committee -from the left, chair Dr. Marten van Schijndel, John Starr,  and committee members Dr. Jennifer Kuo, Dr. Helena Aparicio, and on zoom Dr. Draga Zec.

Abstract:

It is well-established that people construct representations at multiple levels when processing linguistic input: for example, during and after reading "Where is my capybara?", people develop phonological, syntactic, semantic, and discourse representations of the sentence and its components.

The primary focus of this thesis is understanding how people contextually use and prioritize these levels of representations, or streams, during both on-line and off-line processing. Through experiments on three phenomena (phonotactic distinctions, word order preferences for binomials, and rhyme), I show how these streams interface with one another, focusing on how phonological information is incorporated with linguistic structures that are more commonly studied in the sentence processing literature.

Broadly, my findings support a model of sentence processing that allows for dynamic, incremental interactions between levels of linguistic structure.

On March 24, 2026 PhD Candidate Chloe Dokyung Kwon successfully defended her dissertation:   Korean word-medial stops and compound tensification: Acoustic and perceptual studies - see the abstract below. 

Chloe is shown here with her dissertation committee - co-chairs Drs. Sam Tilsen and Abby Cohn, and committee member Dr. John Whitman on Zoom.

Abstract:

This dissertation investigates word-medial stop realization in Korean through the lens of compound tensification — a process whereby a plain stop at the onset of the second noun in a noun-noun compound often surfaces as tense — through a production study of 32 speakers and a perception study of 94 listeners.

Acoustic analyses reveal that derived tense stops converge categorically on underlying tense stops rather than approximating them gradually, and that word-medial stop contrasts are maintained primarily through closure duration rather than the F0 and voice quality cues that predominate word-initially, suggesting a redistribution of acoustic correlates rather than positional weakening or merger. Listeners likewise rely on closure duration when categorizing medial stops; confidence ratings suggest sensitivity to phonological context, though this does not alter cue weighting.

Together, these findings disentangle categoriality and gradience in phonetic realization from variability, demonstrate that individual speaker variation is highly systematic and likely tracked by listeners, and suggest that positional differences in stop realization reflect differences in gestural timing relations rather than weakening or neutralization.

On October 9 & 10, 2025, Dr. Mark Tiede from the Yale School of Medicine conducted an intensive two-day workshop on ultrasonic imaging techniques at the Phonetics Laboratory. The workshop, which combined theoretical instruction with hands-on experimentation, introduced participants to the practical aspects of imaging the human vocal apparatus in real-time. 

The workshop's first day began in the Ultrasound Lab with comprehensive coverage of fundamental scanning parameters, including depth of field, sector angle, and frequency settings. Dr. Tiede demonstrated how these elements influence image quality and frame rates while addressing common challenges such as hyoid and mandibular shadowing.  

The second day  focused on more specialized applications. Topics included probe stabilization techniques using the ALPHUS system, co-collection of ultrasound data with intraoral EMA, and transverse laryngeal imaging for studying devoicing gestures. The workshop concluded with a  pilot experiment in which workshop participants could see their own vocal fold vibrations in real time. 

Everyone had fun and learned quite a bit over the two days, and a number of students will be applying Dr. Tiede's information to their own Phonetics lab ultrasonic imaging experiments in the coming months.